- #1
Diffy
- 441
- 0
Hi, this contracted company that I am working with has this data. Basically they collect the number of transactions each day. Then they graph these transactions out by Week, always monday - sunday.
The data is then stored in a table where the x-axis is the date of the monday the week started and the y-axis is the average of the transactions for that week. They like it that way because it smooths out the weekends.
The problem is they are reporting a growth rate. Since the last couple weeks of the year are generally slower because of holidays, they appear to pick the second or third last full week of the year, and the second or third last week of the previous year, divide, and then come up with a rate.
Given the data, and the structure that they keep it, is there a better way?
I was thinking that they could average the weekly averages of each year. But I remember an old college professor always used to warn against averaging averages.
Any input is appreciated.
The data is then stored in a table where the x-axis is the date of the monday the week started and the y-axis is the average of the transactions for that week. They like it that way because it smooths out the weekends.
The problem is they are reporting a growth rate. Since the last couple weeks of the year are generally slower because of holidays, they appear to pick the second or third last full week of the year, and the second or third last week of the previous year, divide, and then come up with a rate.
Given the data, and the structure that they keep it, is there a better way?
I was thinking that they could average the weekly averages of each year. But I remember an old college professor always used to warn against averaging averages.
Any input is appreciated.